refactor(compressors): switch SAM from automatic mask generation to bbox-prompted segmentation

Browse Source

- Replace SAM2AutomaticMaskGenerator pipeline with Sam2Processor+Sam2Model
- Freeze SAM model parameters at load time, removing torch.no_grad() at call sites
- Rewrite proposal/core.py to use bbox prompts instead of automatic point sampling
- Add bboxes parameter to all HashPipeline public methods (forward, forward_dataset, extract_features, extract_features_dataset)
- Extract mask filtering logic (_filter_masks) from proposal into pipeline
- Rename object_score/ to filter/
- Add load_owlv2_model to model_loader
- Rename notebooks/test.py to habitat_sim_setup.py

...

This commit is contained in:

SikongJueluo

2026-04-02 16:42:59 +08:00

parent eaf02cc97a

commit 42acb3ee1b

9 changed files with 161 additions and 167 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

0

notebooks/test.py → notebooks/habitat_sim_setup.py